jat

🥈Silver

The World's First Agentic IDE. Visual dashboard: live sessions, task management, code editor, terminal. Epic Swarm parallel workflows. Auto-proceed rules. Automation patterns. Beads + Agent Mail + 50 bash tools. Supervise 20+ agents from one UI.

19310Updated 2mo ago

Intermediate30min to implementautomation

Saves ~180 min per use

Quick InstallView Source

git clone https://github.com/joewinke/jat.git

Works with:

Claude

Overview

About This Skill

How to Use

[{"step":"Define the task and subtasks. Break down the automation goal into parallelizable chunks (e.g., 'audit 5 pages for performance, accessibility, and SEO'). Use JAT's swarm feature to assign each subtask to a dedicated agent.","tip":"Label agents clearly (e.g., 'Agent-Performance', 'Agent-Accessibility') and document their roles in the JAT dashboard to avoid confusion."},{"step":"Configure auto-proceed rules. Set conditions for when the swarm should pause (e.g., 'halt if any agent reports a 500 error') or continue (e.g., 'skip re-testing pages with no changes'). Use JAT's rule engine to define these triggers.","tip":"Test auto-proceed rules with a small subset of data first to ensure they work as intended before scaling up."},{"step":"Select and attach tools. Use JAT's built-in tools (e.g., Lighthouse, axe-core) or integrate custom bash scripts. Ensure all agents have the necessary permissions and dependencies.","tip":"Leverage JAT's 'Beads' feature to chain tools together (e.g., run Lighthouse, then parse results with a custom script)."},{"step":"Monitor and review results. Use the JAT dashboard to track agent progress in real-time. Check for errors, warnings, and completed tasks. Export logs for further analysis if needed.","tip":"Use JAT's 'Agent Mail' to send automated email alerts for critical issues (e.g., '503 error detected on /blog')."},{"step":"Iterate and optimize. After the first run, review the summary report to identify bottlenecks or recurring issues. Adjust the swarm configuration, tools, or auto-proceed rules for future runs.","tip":"Save successful configurations as 'Epic Patterns' in JAT for reuse in future projects."}]

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/joewinke/jat

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Use JAT (Just an Agent Thing) to automate the following task: [DESCRIBE_TASK]. Set up a swarm of [NUMBER] agents to handle [SUBTASKS] in parallel. Configure auto-proceed rules for [CONDITIONS] and use [TOOLS] to execute the workflow. Monitor progress in the JAT dashboard and provide a summary report of results. Include any errors or warnings encountered during execution.

Example Output

### JAT Automation Report: Website Performance Audit

**Task:** Conduct a full performance audit of [fictional-site.com](https://fictional-site.com) by testing 12 core pages for load time, accessibility, and SEO issues.

**Swarm Configuration:**
- Deployed 5 agents in parallel (1x Lighthouse Auditor, 1x Accessibility Checker, 1x SEO Validator, 1x Performance Monitor, 1x Error Logger).
- Auto-proceed rules triggered when: (1) All agents complete their initial scan, (2) Any agent reports a critical error (e.g., 404 or 5xx).
- Tools used: Lighthouse CLI, axe-core, Screaming Frog API, Google PageSpeed Insights, and custom bash scripts for log parsing.

**Results Summary:**
- **Total Pages Tested:** 12 (Homepage, About, Blog, 4 Product Pages, Contact, FAQ, Pricing, Login, Cart).
- **Critical Errors Found:** 3 (1x 503 on /blog, 2x broken links in /pricing).
- **Performance Issues:** 8 pages scored <70 on Lighthouse (avg. load time: 4.2s vs. target 2s).
- **Accessibility Violations:** 23 (mostly missing alt text and low-contrast buttons).
- **SEO Issues:** 15 (duplicate meta titles, missing H1s on 3 pages).

**Agent Logs:**
- Performance Monitor flagged /cart as the slowest (6.8s) due to unoptimized images.
- Accessibility Checker identified 15 instances of missing ARIA labels in the checkout flow.
- SEO Validator reported 5 pages with duplicate H1 tags (e.g., /about and /blog both used "Welcome to Fictional Site").

**Next Steps:**
1. Prioritize fixing the 3 critical errors (503 error, broken links) within 24 hours.
2. Schedule a sprint to optimize images in /cart and /pricing (target: reduce load time by 50%).
3. Run a second audit after fixes to verify improvements.

**Dashboard Status:** All agents completed tasks in 12 minutes (vs. estimated 45 minutes sequentially). Errors and warnings are highlighted in red for quick triage.

**Recommendation:** Enable auto-proceed rules for future audits to skip re-testing pages with no changes, saving ~30% time on subsequent runs.